Rapid progress of NGS technologies in clinical practice has led to an unprecedented increase in the investigation of genetic diseases. Many bioinformatics filtering strategies and clinical interpretation guidelines tend to focus on mutations that change the amino acid sequence of proteins. This phenomenon may lead to variants in noncoding regions or synonymous variants being filtered out at the initial analysis. Our limited understanding of the complex genotype-phenotype and a large number of variants hampered our judgment of "benign", “likely benign” or "undetermined significance” variants. The uncertainty of the potential pathogenicity of the variants has confused clinical doctors and genetic data analysts. With the increasing number of detected variants and the expansion of the detection range, diseases may be reclassified after determining the pathogenicity of variants that were originally uncertain or benign. The combination of predictive bioinformatics and functional experiments is a practicable way to uncover the genetic etiology of undiagnosed genetic diseases.
In this study, based on the clinical data and histological characteristics of these siblings, infantile NPHP was established previously [13]. However, at initial analysis, only a likely pathogenic intron variant NG_008130.2(NM_153240.5):c.2088 + 5G > A of NPHP3 has been suggested, and the genetic cause remains undetermined. After reanalyzing WES data of the proband by more predicting tools, we identified an allelic synonymous variant, NG_008130.2(NM_153240.5):c.2154C > T:p.(Phe718Phe) of NPHP3 in trans. We then performed immunofluorescence staining to detect the expression of nephrocystin 3 in the kidney tissue of the proband and found that it was absent. Both variants cause aberrant splicing. RT‒PCR and mRNA sequencing of RNAs from URECs of the carrier of the synonymous variant NG_008130.2(NM_153240.5):c.2154C > T:p.(Phe718Phe) (the mother) and minigene assays of both variants were performed. RT‒PCR of mRNA from the mother with the synonymous variant showed an additional faint short band, which was confirmed to show transcript loss of exon 15 by mRNA sequencing. This result is compatible with a previous study by Elisa Molinari (Molinari, et al., 2018). A minigene assay also showed that the NPHP3 variant NG_008130.2(NM_153240.5):c.2088 + 5G > A led to exon 14 skipping. Therefore, the synonymous variant compounded with the intron variant of NPHP3 explained the previously undiagnosed infantile NPHP of this pedigree.
RNA splicing is an essential process for transferring information from the DNA level to the protein level. During splicing, the noncoding sequences of genes (introns) from the precursor transcript (pre-mRNA) are cut out, and the protein-coding sequences (exons) are joined together to enable translation of messenger RNA into protein. We predicted the effects of these two candidate pathogenic variants on splicing using HSF3.1, BDGP, and SpliceAI. For NPHP3 NG_008130.2(NM_153240.5):c.2088 + 5G > A, all three prediction results strongly suggest that the variant can affect splicing. NPHP3 NG_008130.2(NM_153240.5):
c.2154C > T is located in the 18th nucleotide position of NPHP3 exon 15 and near the exon 15–intron 15 boundary. HSF3.1 was predicted to destroy the exon splicing enhancer (ESE) and gain the exon splicing silencer (ESS). BDGP and Splice AI predicted that it cannot affect splicing (Table 5). In addition, Mutation Taster (https://www.mutationtaster.org/,accessed July 2020) predicted that both variants would lead to protein features and splice site changes that result in the loss of the repeat TPR domain (2 ~ 11) in NPHP3 (downstream of the altered splice site) (Fig. 3B). Different prediction algorithms had the same results for NPHP3 NG_008130.2(NM_153240.5):c.2088 + 5G > A, but they differed in whether NPHP3 NG_008130.2(NM_153240.5):c. 2154C > T had an effect on splicing.
Table 5
Splice prediction of NPHP3 variants by bioinformatics tools
Variant
|
NG_008130.2(NM_153240.5):
c.2088 + 5G > A
|
NG_008130.2(NM_153240.5):
c.2154C > T
|
Location
|
Intron 14 (+ 5)
|
Exon 15 (-18)
|
5' Splice site
|
GCAGgtgtgtatttttc
GCAGgtgtatatttttc
|
ACCCTTTTCGGCAAAAT
ACCCTTTTTGGCAAAAT
|
HSF3.1
|
Broken WT Donor Site
89.96–80.4 (-10.63%)
MaxEntScan
6.99–3.09 (-55.79%)
|
Broken ESE site and gain new ESS
|
BDGP
|
WT Donor Site 0.95
MUT Donor Site 0.42
|
WT Donor Site 0.96
MUT Donor Site 0.96
|
Splice AI
|
WT Donor Loss 0.95
WT Acceptor Loss 0.01
|
WT Donor Loss 0.01
|
Notes: ESE = exon splicing enhancer; ESS = exon splicing silencer |
However, in silico algorithms do not always predict the exact splicing abnormalities. If the donor or acceptor site in internal exons is abnormal or a new splicing site is produced, the exons may be lost or truncated, and the intron may become part of the exon. In this scenario, the splicing site variation will not always be predicted to be positive [10]. Therefore, whether this type of variant has an effect on protein function needs to be verified. The simplest and most effective method to analyze RNA directly is using human specimens. However, RNA samples from the affected tissue of patients are sometimes difficult to obtain, and even when they are available, detection may be challenging because of RNA degradation. RNA samples of these family members were not available initially in this study, so we investigated the potential pathogenicity of these specific splice variants using a classic splicing minigene assay, which is a powerful tool for studying abnormally spliced SNPs. The results of minigene experiments in vitro showed that NPHP3 NG_008130.2(NM_153240.5):c.2088 + 5G > A led to whole exon 14 skipping. However, NPHP3 NG_008130.2(NM_153240.5):c.2154C > T:p.(Phe718Phe)variant causing an exon 15-skipping transcript, resulted in an increase to approximately 65% compared with the control in both cell lines. Both exon 15 skipping and exon 15-including transcripts were detected in WT lines, as well as in c.2154C > T variant cell lines. There was a higher ratio of exon 15-included transcripts in WT and a higher ratio of exon 15 skipping transcripts in c.2154C > T variant cell lines. These in vitro results are different from those of URECs in vivo in this study and in Elisa Molinari's study (Molinari, et al., 2018). In Elisa Molinari's study, the researchers further demonstrated that RNA transcripts from peripheral blood white cells showed complete splicing out of exon 15 in healthy controls. These results indicated that mRNA splicing is tissue specific and may be affected by the environment, and minigene assay results may not be faithfully consistent with those in vivo.
Synonymous mutations were previously thought to be "silent mutations" since they cannot affect the amino acid sequence of the coding protein. Nonetheless, a growing amount of evidence suggests that their impact on splicing, RNA stability, RNA translation or cotranslational protein folding can lead to abnormal protein expression levels, structure, and function [11]. The importance of synonymous variants in the genes underlying polycystic kidney disease and NPHP has been described [1; 2; 7; 9], as well as the results in this study.
In conclusion, we verified the pathogenicity of a synonymous variant of NPHP3 in a Chinese infantile NPHP pedigree. These findings suggest that we need to reassess the pathogenicity of candidate synonymous mutations, especially if they are highly consistent with the phenotype. We should consider that minigene assay results may not be faithfully consistent with those in vivo.